Abstract

Scholars who consider that the Transcendental Analytic contains the core of what Kant calls ‘transcendental idealism’ are mistaken. Indeed, Kant’s transcendental idealism of space, time and spatiotemporal objects is sufficiently proved in the Transcendental Aesthetic and does not depend on complementary claims made later on in the Critique. This does not mean, however, that we are allowed to subscribe to the so-called separability-thesis, which states that we can endorse Kant's views in the Transcendental Logic without endorsing the results of the expositions and arguments laid out in the Aesthetic. Nor does it mean that the Aesthetic contains the only proof for transcendental idealism. Indeed, as Kant himself explicitly recognizes, the antinomies provide the premise for an indirect or mediate proof of his position with respect to the nature of space and time. First, I analyze Kant’s argument for the claim that the production of transcendental illusion is inherent to the function of human reasoning, and therefore inevitable. I then follow the presuppositions and the argumentation of the mediate proof of transcendental idealism and argue that it has a more general validity than could be assumed on the basis of Kant’s text, because it is in its essence independent from the particular determinations of the cosmological idea in the form of the four antinomies. Indeed, as I argue, the nature of reason itself and the content of the general cosmological idea suffice as elements for a mediate proof. Thus, I claim that transcendental idealism can be mediately established on the basis of the general antinomical relation between transcendental realism and material idealism, without any requirement to refer to the particular realisations of this antagonism as they are presented in the four antinomies. If the argument is correct, it shows that the mediate proof of transcendental idealism does not depend on a sympathetic account of the antinomies and is not automatically invalidated if the proofs of thesis and antithesis are considered unconvincing.